Predicting Customer Churn Through
Multiple Logistic Regression: A
Complete Performance Assessment
Guide for WGU D600 Task 2 —
Featuring Python Implementation,
Confusion Matrix Evaluation, and
ROC/AUC Analysis (2026 Edition)"
SECTION 1: LOGISTIC REGRESSION
FUNDAMENTALS
Questions 1–25
1. What is the primary purpose of logistic regression?
A. Predict a continuous numerical value
B. Predict the probability of a categorical outcome
C. Calculate the mean of a dataset
D. Reduce the number of observations
Answer: B
Rationale: Logistic regression models the probability of an outcome, commonly a binary
outcome such as churn versus no churn.
2. In a binary logistic regression model, the dependent variable usually has how many
categories?
A. One
B. Two
C. Three
D. Four
Answer: B
Rationale: Binary logistic regression is designed for outcomes with two possible categories,
such as Yes/No.
,3. Which variable would be an appropriate dependent variable for a customer churn
model?
A. Monthly charges
B. Customer age
C. Churn status
D. Contract length
Answer: C
Rationale: Churn status is the outcome being predicted, typically represented as 0 = did not
churn and 1 = churned.
4. What does a predicted probability of 0.80 indicate in a churn model?
A. The customer definitely churned
B. The customer has an estimated 80% probability of churning
C. The model is 80% accurate
D. The customer has an 80% monthly charge
Answer: B
Rationale: A predicted probability represents the model's estimated likelihood of belonging
to the positive class.
5. Which function is commonly used to convert logistic regression scores into
probabilities?
A. Linear function
B. Sigmoid function
C. Square-root function
D. Log function only
Answer: B
Rationale: The sigmoid, or logistic, function converts the linear predictor into a value
between 0 and 1.
6. Why is logistic regression preferable to ordinary linear regression for a binary
outcome?
,A. Logistic regression restricts predicted probabilities to 0–1
B. Logistic regression always has higher accuracy
C. Linear regression cannot use numerical predictors
D. Logistic regression requires no data preparation
Answer: A
Rationale: Probabilities must fall between 0 and 1, and the logistic function naturally
constrains predictions to this range.
7. What does an odds ratio greater than 1 generally indicate?
A. Decreased odds of the outcome
B. Increased odds of the outcome
C. No relationship
D. Missing data
Answer: B
Rationale: An odds ratio above 1 means that an increase in the predictor is associated with
increased odds of the modeled outcome, holding other predictors constant.
8. What does an odds ratio of 1 indicate?
A. The predictor has no change in the odds of the outcome
B. The predictor guarantees the outcome
C. The model is invalid
D. The outcome is impossible
Answer: A
Rationale: An odds ratio of exactly 1 indicates no change in odds associated with a one-unit
increase in the predictor.
9. What does an odds ratio below 1 generally suggest?
A. Increased odds
B. Decreased odds
C. Perfect prediction
D. No observations
Answer: B
, Rationale: An odds ratio less than 1 indicates lower odds of the outcome as the predictor
increases, assuming other variables remain constant.
10. What is the logit?
A. The logarithm of the odds
B. The logarithm of the probability only
C. The square of the probability
D. The average of the predictors
Answer: A
Rationale: The logit is defined as the natural logarithm of the odds of the event.
11. Which is an example of a binary categorical outcome?
A. Salary
B. Age
C. Churn: Yes/No
D. Monthly spending
Answer: C
Rationale: Yes/No contains two possible categories and is therefore appropriate for binary
logistic regression.
12. Multiple logistic regression differs from simple logistic regression because it:
A. Uses multiple observations
B. Uses multiple predictors
C. Uses multiple dependent variables
D. Requires multiple datasets
Answer: B
Rationale: Multiple logistic regression estimates the relationship between a binary outcome
and two or more predictor variables.
13. Which is an example of a continuous predictor?
A. Gender category
B. Contract type
Multiple Logistic Regression: A
Complete Performance Assessment
Guide for WGU D600 Task 2 —
Featuring Python Implementation,
Confusion Matrix Evaluation, and
ROC/AUC Analysis (2026 Edition)"
SECTION 1: LOGISTIC REGRESSION
FUNDAMENTALS
Questions 1–25
1. What is the primary purpose of logistic regression?
A. Predict a continuous numerical value
B. Predict the probability of a categorical outcome
C. Calculate the mean of a dataset
D. Reduce the number of observations
Answer: B
Rationale: Logistic regression models the probability of an outcome, commonly a binary
outcome such as churn versus no churn.
2. In a binary logistic regression model, the dependent variable usually has how many
categories?
A. One
B. Two
C. Three
D. Four
Answer: B
Rationale: Binary logistic regression is designed for outcomes with two possible categories,
such as Yes/No.
,3. Which variable would be an appropriate dependent variable for a customer churn
model?
A. Monthly charges
B. Customer age
C. Churn status
D. Contract length
Answer: C
Rationale: Churn status is the outcome being predicted, typically represented as 0 = did not
churn and 1 = churned.
4. What does a predicted probability of 0.80 indicate in a churn model?
A. The customer definitely churned
B. The customer has an estimated 80% probability of churning
C. The model is 80% accurate
D. The customer has an 80% monthly charge
Answer: B
Rationale: A predicted probability represents the model's estimated likelihood of belonging
to the positive class.
5. Which function is commonly used to convert logistic regression scores into
probabilities?
A. Linear function
B. Sigmoid function
C. Square-root function
D. Log function only
Answer: B
Rationale: The sigmoid, or logistic, function converts the linear predictor into a value
between 0 and 1.
6. Why is logistic regression preferable to ordinary linear regression for a binary
outcome?
,A. Logistic regression restricts predicted probabilities to 0–1
B. Logistic regression always has higher accuracy
C. Linear regression cannot use numerical predictors
D. Logistic regression requires no data preparation
Answer: A
Rationale: Probabilities must fall between 0 and 1, and the logistic function naturally
constrains predictions to this range.
7. What does an odds ratio greater than 1 generally indicate?
A. Decreased odds of the outcome
B. Increased odds of the outcome
C. No relationship
D. Missing data
Answer: B
Rationale: An odds ratio above 1 means that an increase in the predictor is associated with
increased odds of the modeled outcome, holding other predictors constant.
8. What does an odds ratio of 1 indicate?
A. The predictor has no change in the odds of the outcome
B. The predictor guarantees the outcome
C. The model is invalid
D. The outcome is impossible
Answer: A
Rationale: An odds ratio of exactly 1 indicates no change in odds associated with a one-unit
increase in the predictor.
9. What does an odds ratio below 1 generally suggest?
A. Increased odds
B. Decreased odds
C. Perfect prediction
D. No observations
Answer: B
, Rationale: An odds ratio less than 1 indicates lower odds of the outcome as the predictor
increases, assuming other variables remain constant.
10. What is the logit?
A. The logarithm of the odds
B. The logarithm of the probability only
C. The square of the probability
D. The average of the predictors
Answer: A
Rationale: The logit is defined as the natural logarithm of the odds of the event.
11. Which is an example of a binary categorical outcome?
A. Salary
B. Age
C. Churn: Yes/No
D. Monthly spending
Answer: C
Rationale: Yes/No contains two possible categories and is therefore appropriate for binary
logistic regression.
12. Multiple logistic regression differs from simple logistic regression because it:
A. Uses multiple observations
B. Uses multiple predictors
C. Uses multiple dependent variables
D. Requires multiple datasets
Answer: B
Rationale: Multiple logistic regression estimates the relationship between a binary outcome
and two or more predictor variables.
13. Which is an example of a continuous predictor?
A. Gender category
B. Contract type
